Vous êtes sur la page 1sur 12

Click to edit Master subtitle style

3/17/12

OLAP Vs OLTP

Two commonly used acronyms in the database world are OLTP and OLAP. OLTP stands for "Online Transactional Processing" and OLAP represents "Online Analytical Processing." The purposes of the two systems are vastly different, leading to different approaches in organizing and storing the data. OLTP (On-line Transaction Processing) is characterized by a large number of short on-line transactions (INSERT, UPDATE, DELETE). The main emphasis for OLTP systems is put on very fast query processing, maintaining data integrity in multi-access environments and an effectiveness measured by number of transactions per second. In OLTP database there is detailed and current data, and schema used to store transactional databases is the entity model (usually 3NF).
3/17/12

OLAP (On-line Analytical Processing) is characterized by relatively low volume of transactions. Queries are often very complex and involve aggregations. For OLAP systems a response time is an effectiveness measure. OLAP applications are widely used by Data Mining techniques. In OLAP database there is aggregated, historical data, stored in multi-dimensional schemas (usually star schema).

3/17/12

We can divide IT systems into transactional (OLTP) and analytical (OLAP). In general we can assume that OLTP systems provide source data to data warehouses, whereas OLAP systems help to analyze it.

3/17/12

Following is a simple comparison between the two approaches. PURPOSE OLTP: Main purpose is to automate operational procedures and provide up-to-the minute information on current business processes. Reporting tends to be workflow related, answering queries such as "which claims are still open?" and "which customers are waiting for a response? OLAP: Main purpose of this is for reporting and data analysis. This is where users go to answer queries related to business cause, important trends, and overall profitability. Reports from an OLAP are intended to help with general business decision making and future planning, answering queries such as "which demographic attribute makes customers less likely to submit a claim?" or "which insurance policies are most likely to be purchased by those customers who purchase a boat policy?".
3/17/12

DATA SOURCE

OLTP: Data is input by various clerks and office staff, maintaining an operational log. There can be hundreds of users inputting data simultaneously, or there may be only one user. Regardless of the size, data is in a constant state of change. OLAP: Data is rolled in from various data sources, most of which are likely OLTPs that are used in several difference business offices. It may also contain data from flat files or data mined from the web or other businesses. Regardless of the sources, data tends to be slowly updated in a scheduled manner, such as nightly or weekly, allowing statistics to be more easily gathered than in the dynamic OLTP.

3/17/12

IMPORTANT METRICS

OLTP: In OLTP, throughput (number of transactions per second) is vitally important, as is minimizing concurrency conflicts. The point is to allow users to insert and update data quickly so that their daily jobs are not disrupted. This means data organization and indexing should be done in a way that increases the insert and update speed, without being overly concerned about the speed of reporting. OLAP: In OLAP, response time and query throughput are the most important metrics. Since the main purpose is reporting, users expect to be able to gather statistics quickly. Therefore, indexing and data organization may be significantly different from the OLTP, increasing large query response time while not optimizing for inserts and updates.

3/17/12

DATA ORGANIZATION

OLTP: In relational OLTP systems, the data is generally in third-normal form. This lessens the amount of space required and allows fast inserts and updates on data that is logically interrelated. For example, one could update just an event that occurred on a claim, without updating the entire claim itself. OLAP: In relational OLAP systems, the data will usually be in some sort of denormalized form, often in a star schema format based on "facts" and "dimensions." The flattened schema allows the system to retrieve data by going to fewer tables with clear relationships, which increases the speed of complex queries.

3/17/12

INDEXING

OLTP: Indexing can slow down inserts and updates because both the table and the index have to be updated. Therefore, OLTP systems tend to have only a few indexes per table, representing the most common searches related to workflow reports. OLAP: When done right, indexing significantly increases the speed of large queries. It is not uncommon for an OLAP table to have an index on every single column.

3/17/12

PHYSICAL ARCHITECTURE

OLTP: Most OLTP databases are in a relational database form, organizing information according to tables that are connected by a common column, or "key. OLAP: Some OLAP databases may be stored in a relational form, and thus are called ROLAP. Others may be stored in a multi-dimensional engine which stores data in multidimensional arrays. These are called MOLAP. DISK SPACE

OLTP: OLTPs are smaller, generally between a few hundred megabytes and a hundred gigabytes. This is because the third-normal form organization reduces data redundancy and space-intensive indexes are kept to a minimum.
3/17/12

OLAP:OLAPs can be much large, especially since data may come from multiple OLTPs. The denormalized data and extensive indexing also takes up quite a bit of space. OLAPs tend to run in the hundreds of gigabytes into the terabyte range. In summary, OLAP and OLTP are similar in that they are terms used to describe storage of interrelated data. But because the intended use of the systems are different, the physical and logical architecture need to be different in order to optimize performance.

3/17/12

Thank You

3/17/12

Vous aimerez peut-être aussi